Page 1 of 1
SMP and GPU 2 client
Posted: Tue Aug 05, 2008 4:09 pm
by Bikerz
I run a GPU client on my GTX 8800 and get about 5100ppd, then if I turn my CPU smp client on I still only get combined from both about 5200ppd. So no point as far as I can see of running both. This I know is down to my GTX taking 100% of one of my 4 cores. Is there a way to make the SMP client only use 3 cores and the gpu2 client to use the other one? Is that the best thing to do to maximize ppd?
Cheers
Sheldon
Re: SMP and GPU 2 client
Posted: Tue Aug 05, 2008 4:14 pm
by jrweiss
Set the priority for the GPU client to low instead of idle. That should let the GPU client have as much as it needs of 1 core, and allow the SMP client to have the other 3+.
Re: SMP and GPU 2 client
Posted: Tue Aug 05, 2008 4:23 pm
by DriveEuro
It also helps to run Vista for GPU2. Vista requires a lot less CPU cycles for nvidia GPU2 than XP.
Re: SMP and GPU 2 client
Posted: Tue Aug 05, 2008 4:46 pm
by Foxery
Are you sure you have an actual SMP work unit running? (-smp flag set)
Re: SMP and GPU 2 client
Posted: Tue Aug 05, 2008 9:50 pm
by Bikerz
How would I check? Seems to be running ok, but I do admit I dont understand that bit that well. Im on xp 32 bit. How do you change priority on these new clients, I dont know now they have changed?
Re: SMP and GPU 2 client
Posted: Wed Aug 06, 2008 2:57 am
by codysluder
The priority setting is still the same: you have a choice of Lowest or Slightly Higher.
There are two ways to know if you're running a smp WU.
* Are there four copies of FahCore_a1 running?
* Check the project number on the Project Summary page.
Re: SMP and GPU 2 client
Posted: Wed Aug 06, 2008 4:57 am
by jrweiss
Bikerz wrote:How would I check? Seems to be running ok, but I do admit I dont understand that bit that well. Im on xp 32 bit. How do you change priority on these new clients, I dont know now they have changed?
Check the log. When the client first starts, it echos the switches you have set. You can also tell by the WU if it is SMP or single-CPU. Post your log, and we can help.
Priority is set in the Config in the advanced properties.
Re: SMP and GPU 2 client
Posted: Wed Aug 06, 2008 3:32 pm
by Bikerz
heers my SMP client is:
Code: Select all
--- Opening Log file [August 5 15:56:48 UTC]
# Windows SMP Console Edition #################################################
###############################################################################
Folding@Home Client Version 6.22 SMP Beta2
http://folding.stanford.edu
###############################################################################
###############################################################################
Launch directory: C:\Program Files\Folding@Home Windows SMP Client V1.01
Executable: C:\Program Files\Folding@Home Windows SMP Client V1.01\FAH6.22beta2-SMP-shm.exe
Arguments: -smp
[15:56:48] Configuring Folding@Home...
[15:57:21] - Ask before connecting: No
[15:57:21] - User name: Bikerz (Team 35947)
[15:57:21] - User ID: 145DA8768A938CA
[15:57:21] - Machine ID: 1
[15:57:21]
[15:57:21] Work directory not found. Creating...
[15:57:21] Could not open work queue, generating new queue...
[15:57:21] - Preparing to get new work unit...
[15:57:21] + Attempting to get work packet
[15:57:21] - Connecting to assignment server
[15:57:25] - Successful: assigned to (171.64.65.64).
[15:57:25] + News From Folding@Home: Welcome to Folding@Home
[15:57:26] Loaded queue successfully.
[16:04:16] + Closed connections
[16:04:16]
[16:04:16] + Processing work unit
[16:04:16] Work type a1 not eligible for variable processors
[16:04:16] Core required: FahCore_a1.exe
[16:04:16] Core not found.
[16:04:16] - Core is not present or corrupted.
[16:04:16] - Attempting to download new core...
[16:04:16] + Downloading new core: FahCore_a1.exe
[16:04:18] + 10240 bytes downloaded
...snip...
[16:05:26] + 789667 bytes downloaded
[16:05:26] Verifying core Core_a1.fah...
[16:05:26] Signature is VALID
[16:05:26]
[16:05:26] Trying to unzip core FahCore_a1.exe
[16:05:26] Decompressed FahCore_a1.exe (2035712 bytes) successfully
[16:05:31] + Core successfully engaged
[16:05:36]
[16:05:36] + Processing work unit
[16:05:36] Work type a1 not eligible for variable processors
[16:05:36] Core required: FahCore_a1.exe
[16:05:36] Core found.
[16:05:36] Using generic mpiexec calls
[16:05:36] Working on queue slot 01 [August 5 16:05:36 UTC]
[16:05:36] + Working ...
[16:05:36]
[16:05:36] *------------------------------*
[16:05:36] Folding@Home Gromacs SMP Core
[16:05:36] Version 1.74 (March 10, 2007)
[16:05:36]
[16:05:36] Preparing to commence simulation
[16:05:36] - Looking at optimizations...
[16:05:37] - Created dyn
[16:05:37] - Files status OK
[16:05:51] - Expanded 4715335 -> 24426905 (decompressed 518.0 percent)
[16:05:52] - Starting from initial work packet
[16:05:52]
[16:05:52] Project: 2665 (Run 3, Clone 208, Gen 32)
[16:05:52]
[16:05:52] Assembly optimizations on if available.
[16:05:52] Entering M.D.
[16:06:00] Rejecting checkpoint
[16:06:02] Protein: HGG in water
[16:06:02] Writing local files
[16:06:11] Extra SSE boost OK.
[16:06:12] Writing local files
[16:06:12] Completed 0 out of 250000 steps (0 percent)
[16:24:44] Writing local files
...snip...
[21:52:15] Completed 47500 out of 250000 steps (19 percent)
Folding@Home Client Shutdown at user request.
Folding@Home Client Shutdown.
AND GPU client is:
--- Opening Log file [August 4 16:21:00 UTC]
# Windows GPU Console Edition #################################################
###############################################################################
Folding@Home Client Version 6.20
http://folding.stanford.edu
###############################################################################
###############################################################################
Launch directory: C:\Documents and Settings\Admin\Desktop\Folding\GPU
Executable: C:\Documents and Settings\Admin\Desktop\Folding\GPU\Folding@home-Win32-GPU.exe
[16:21:00] Configuring Folding@Home...
[16:21:30] - Ask before connecting: No
[16:21:30] - User name: Bikerz (Team 35947)
[16:21:30] - User ID: 145DA8768A938CA
[16:21:30] - Machine ID: 2
[16:21:30]
[16:21:30] Work directory not found. Creating...
[16:21:30] Could not open work queue, generating new queue...
[16:21:30] - Preparing to get new work unit...
[16:21:30] + Attempting to get work packet
[16:21:30] - Connecting to assignment server
[16:21:35] - Successful: assigned to (171.64.65.106).
[16:21:35] + News From Folding@Home: GPU folding beta
[16:21:35] Loaded queue successfully.
[16:21:41] + Closed connections
[16:21:41]
[16:21:41] + Processing work unit
[16:21:41] Core required: FahCore_11.exe
[16:21:41] Core not found.
[16:21:41] - Core is not present or corrupted.
[16:21:41] - Attempting to download new core...
[16:21:41] + Downloading new core: FahCore_11.exe
[16:21:44] + 10240 bytes downloaded
...snip...
[16:22:25] Verifying core Core_11.fah...
[16:22:25] Signature is VALID
[16:22:25]
[16:22:25] Trying to unzip core FahCore_11.exe
[16:22:25] Decompressed FahCore_11.exe (1822720 bytes) successfully
[16:22:30] + Core successfully engaged
[16:22:35]
[16:22:35] + Processing work unit
[16:22:35] Core required: FahCore_11.exe
[16:22:35] Core found.
[16:22:35] Working on queue slot 01 [August 4 16:22:35 UTC]
[16:22:35] + Working ...
[16:22:35]
[16:22:35] *------------------------------*
[16:22:35] Folding@Home GPU Core - Beta
[16:22:35] Version 1.09 (Fri Aug 1 11:46:54 PDT 2008)
[16:22:35]
[16:22:35] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[16:22:35] Build host: amoeba
[16:22:35] Board Type: Nvidia
[16:22:35] Core :
[16:22:35] Preparing to commence simulation
[16:22:35] - Looking at optimizations...
[16:22:35] - Created dyn
[16:22:35] - Files status OK
[16:22:35] - Expanded 39220 -> 229213 (decompressed 584.4 percent)
[16:22:35] Called DecompressByteArray: compressed_data_size=39220 data_size=229213, decompressed_data_size=229213 diff=0
[16:22:35] - Digital signature verified
[16:22:35]
[16:22:35] Project: 5505 (Run 0, Clone 64, Gen 0)
[16:22:35]
[16:22:35] Assembly optimizations on if available.
[16:22:35] Entering M.D.
[16:22:42] Working on 544 p5505_fip35_ww_domain
[16:22:42] Client config found, loading data.
[16:22:42] Starting GUI Server
[16:24:10] Completed 1%
...snip...
[19:13:14] Completed 100%
[19:14:14]
[19:14:14] Finished Work Unit:
[19:14:14] - Reading up to 32880 from "work/wudata_01.trr": Read 32880
[19:14:14] trr file hash check passed.
[19:14:14] - Reading up to 534028 from "work/wudata_01.xtc": Read 534028
[19:14:14] xtc file hash check passed.
[19:14:14] edr file hash check passed.
[19:14:14] logfile size: 78207
[19:14:14] Leaving Run
[19:14:16] - Writing 646187 bytes of core data to disk...
[19:14:16] Done: 645675 -> 568306 (compressed to 88.0 percent)
[19:14:16] ... Done.
[19:14:16] - Shutting down core
[19:14:16]
[19:14:16] Folding@home Core Shutdown: FINISHED_UNIT
[19:14:20] CoreStatus = 64 (100)
[19:14:20] Sending work to server
[19:14:20] Project: 5505 (Run 0, Clone 64, Gen 0)
[19:14:20] - Read packet limit of 540015616... Set to 524286976.
[19:14:20] + Attempting to send results [August 4 19:14:20 UTC]
[19:14:47] + Results successfully sent
[19:14:47] Thank you for your contribution to Folding@Home.
[19:14:47] + Starting local stats count at 1
[19:14:51] - Preparing to get new work unit...
[19:14:51] + Attempting to get work packet
[19:14:51] - Connecting to assignment server
[19:14:56] - Successful: assigned to (171.64.65.20).
[19:14:56] + News From Folding@Home: GPU folding beta
[19:14:56] Loaded queue successfully.
[19:15:16] + Closed connections
[19:15:16]
[19:15:16] + Processing work unit
[19:15:16] Core required: FahCore_11.exe
[19:15:16] Core found.
[19:15:16] Working on queue slot 02 [August 4 19:15:16 UTC]
[19:15:16] + Working ...
[19:15:16]
[19:15:16] *------------------------------*
[19:15:16] Folding@Home GPU Core - Beta
[19:15:16] Version 1.09 (Fri Aug 1 11:46:54 PDT 2008)
[19:15:16]
[19:15:16] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[19:15:16] Build host: amoeba
[19:15:16] Board Type: Nvidia
[19:15:16] Core :
[19:15:16] Preparing to commence simulation
[19:15:16] - Looking at optimizations...
[19:15:16] - Created dyn
[19:15:16] - Files status OK
[19:15:16] - Expanded 43799 -> 244433 (decompressed 558.0 percent)
[19:15:16] Called DecompressByteArray: compressed_data_size=43799 data_size=244433, decompressed_data_size=244433 diff=0
[19:15:16] - Digital signature verified
[19:15:16]
[19:15:16] Project: 5008 (Run 1, Clone 93, Gen 73)
[19:15:16]
[19:15:16] Assembly optimizations on if available.
[19:15:16] Entering M.D.
[19:15:22] Working on 576 p5005_supervillin_e1
[19:15:23] Client config found, loading data.
[19:15:23] Starting GUI Server
[19:17:48] Completed 1%
...snip...
[22:01:23] Completed 68%
Folding@Home Client Shutdown.
--- Opening Log file [August 5 16:02:23 UTC]
# Windows GPU Console Edition #################################################
###############################################################################
Folding@Home Client Version 6.20
http://folding.stanford.edu
###############################################################################
###############################################################################
Launch directory: C:\Documents and Settings\Admin\Desktop\Folding\GPU
Executable: C:\Documents and Settings\Admin\Desktop\Folding\GPU\Folding@home-Win32-GPU.exe
[16:02:23] - Ask before connecting: No
[16:02:23] - User name: Bikerz (Team 35947)
[16:02:23] - User ID: 145DA8768A938CA
[16:02:23] - Machine ID: 2
[16:02:23]
[16:02:23] Loaded queue successfully.
[16:02:23]
[16:02:23] + Processing work unit
[16:02:23] Core required: FahCore_11.exe
[16:02:23] Core found.
[16:02:23] Working on queue slot 02 [August 5 16:02:23 UTC]
[16:02:23] + Working ...
[16:02:23]
[16:02:23] *------------------------------*
[16:02:23] Folding@Home GPU Core - Beta
[16:02:23] Version 1.09 (Fri Aug 1 11:46:54 PDT 2008)
[16:02:23]
[16:02:23] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[16:02:23] Build host: amoeba
[16:02:23] Board Type: Nvidia
[16:02:23] Core :
[16:02:23] Preparing to commence simulation
[16:02:23] - Looking at optimizations...
[16:02:23] - Files status OK
[16:02:23] - Expanded 43799 -> 244433 (decompressed 558.0 percent)
[16:02:23] Called DecompressByteArray: compressed_data_size=43799 data_size=244433, decompressed_data_size=244433 diff=0
[16:02:23] - Digital signature verified
[16:02:23]
[16:02:23] Project: 5008 (Run 1, Clone 93, Gen 73)
[16:02:23]
[16:02:23] Assembly optimizations on if available.
[16:02:23] Entering M.D.
[16:02:30] Will resume from checkpoint file
[16:02:30] Working on 576 p5005_supervillin_e1
[16:02:31] Client config found, loading data.
[16:02:31] Starting GUI Server
[16:02:31] Resuming from checkpoint
[16:02:31] Verified work/wudata_02.log
[16:02:31] Verified work/wudata_02.edr
[16:02:31] Verified work/wudata_02.trr
[16:02:31] Verified work/wudata_02.xtc
[16:02:31] Completed 68%
...snip...
[17:40:23] Completed 100%
[17:41:23]
[17:41:23] Finished Work Unit:
[17:41:23] - Reading up to 34800 from "work/wudata_02.trr": Read 34800
[17:41:23] trr file hash check passed.
[17:41:23] - Reading up to 1130600 from "work/wudata_02.xtc": Read 1130600
[17:41:23] xtc file hash check passed.
[17:41:23] edr file hash check passed.
[17:41:23] logfile size: 130667
[17:41:23] Leaving Run
[17:41:26] - Writing 1297139 bytes of core data to disk...
[17:41:26] Done: 1296627 -> 1159666 (compressed to 89.4 percent)
[17:41:26] ... Done.
[17:41:27] - Shutting down core
[17:41:27]
[17:41:27] Folding@home Core Shutdown: FINISHED_UNIT
[17:41:29] CoreStatus = 64 (100)
[17:41:29] Sending work to server
[17:41:29] Project: 5008 (Run 1, Clone 93, Gen 73)
[17:41:29] - Read packet limit of 540015616... Set to 524286976.
[17:41:29] + Attempting to send results [August 5 17:41:29 UTC]
[17:44:53] + Results successfully sent
[17:44:53] Thank you for your contribution to Folding@Home.
[17:44:53] + Number of Units Completed: 2
[17:44:57] - Preparing to get new work unit...
[17:44:57] + Attempting to get work packet
[17:44:57] - Connecting to assignment server
[17:45:07] - Successful: assigned to (171.64.65.106).
[17:45:07] + News From Folding@Home: GPU folding beta
[17:45:07] Loaded queue successfully.
[17:45:13] + Closed connections
[17:45:13]
[17:45:13] + Processing work unit
[17:45:13] Core required: FahCore_11.exe
[17:45:13] Core found.
[17:45:13] Working on queue slot 03 [August 5 17:45:13 UTC]
[17:45:13] + Working ...
[17:45:13]
[17:45:13] *------------------------------*
[17:45:13] Folding@Home GPU Core - Beta
[17:45:13] Version 1.09 (Fri Aug 1 11:46:54 PDT 2008)
[17:45:13]
[17:45:13] Compiler : Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 14.00.50727.762 for 80x86
[17:45:13] Build host: amoeba
[17:45:13] Board Type: Nvidia
[17:45:13] Core :
[17:45:13] Preparing to commence simulation
[17:45:13] - Looking at optimizations...
[17:45:13] - Created dyn
[17:45:13] - Files status OK
[17:45:13] - Expanded 45464 -> 246249 (decompressed 541.6 percent)
[17:45:13] Called DecompressByteArray: compressed_data_size=45464 data_size=246249, decompressed_data_size=246249 diff=0
[17:45:13] - Digital signature verified
[17:45:13]
[17:45:13] Project: 5506 (Run 9, Clone 183, Gen 5)
[17:45:13]
[17:45:13] Assembly optimizations on if available.
[17:45:13] Entering M.D.
[17:45:20] Working on p5506_supervillin_e1
[17:45:20] Client config found, loading data.
[17:45:20] Starting GUI Server
[17:48:23] Completed 1%
...snip...
[21:46:40] Completed 77%
Folding@Home Client Shutdown.
Id like to know how to increase my PPD, by setting gpu to use 1 core and the smp to use the other 3?
PM Sent. -7im
Re: SMP and GPU 2 client
Posted: Wed Aug 06, 2008 4:26 pm
by Flathead74
Bikerz, you might try
Set Affinity II - v1.036, it works for me.
Re: SMP and GPU 2 client
Posted: Wed Aug 06, 2008 5:14 pm
by jrweiss
Id like to know how to increase my PPD, by setting gpu to use 1 core and the smp to use the other 3?
It will do essentially that by setting the priority in the GPU client to Low (or "Slightly higher", depending on the client version). In the console client, shut down the GPU client and run it in a Command Prompt window with the
-configonly switch. In the advanced properties, adjust the priority. In the systray version, use the config option and then restart.
Re: SMP and GPU 2 client
Posted: Wed Aug 06, 2008 11:37 pm
by Jeannie
Based on your log, you're running a Project 2665 work unit, and that's definitely an SMP work unit, in case you still had a question about that.